Towards Behavior Trees Based Robotic Task Execution for Physical Human Robot Collaboration

This work presents an innovative framework for enhancing human-robot collaboration in industrial settings through Behavior Tree-based Robot Task Execution (BTE) and advanced task planning. The BTE framework orchestrates robot skills in assembly scenarios, ensuring smooth task execution and workflow synchronization. Concurrently, the task planning system integrates diverse sensing channels—such as voice, force feedback, and graphical user interfaces—to accurately discern human intentions for collaborative tool handovers. A significant novelty of our approach is its emphasis on human-centric factors, including ergonomics and user preferences, to optimize tool handover position and timing. By synergizing task planning, execution, and redundant sensing, our framework enables seamless human robot collaboration. The key contributions include a structured BTE framework for robot task execution and advanced task planning, integrating redundant sensing for intuitive collaboration. Real-world experiments in car door assembly validate the practicality and adaptability of our approach, emphasizing efficiency and adaptability in human-robot collaboration.